Calculating accurate conductance across model systems to develop techniques for transport across single-molecular junctions.
The current in a molecular junction is a response property (at weak bias) and requires a non-equilibrium treatment (at finite bias), traditional ab initio methods of quantum chemistry and of ground-state DFT have long been regarded as insufficient. There are many well-known approaches that have been developed in many-body physics from decades of studying this problem for mesoscopic systems, such as quantum dots. Such methods are often so computationally demanding that they can only be applied to simplified Hamiltonians, such as the Anderson model or a Hubbard chain. Thus they are not first-principles and do not produce chemically realistic results, or do so only in an empirical fashion. To fill this gap, a standard approach for performing such calculations within DFT was developed early on and is often referred to as non-equilibrium Green’s function (NEGF). In this model, a ground-state DFT calculation is performed for the system with a bias applied, and then the current through the ground-state KS potential is calculated via the Landauer formalism. The Landauer approach can be derived from non-equilibrium Green’s functions, scattering theory, or Kubo linear response. This generally works well for both metal wires and carbon nanotubes, where conductance is simply the product of the number of open channels times the fundamental unit of conductance. However, in the technologically important area of organic molecules between metal leads, standard model calculations often yield conductances that are one or two orders of magnitude larger than experiment.
We endeavor to further improve and understand the modern methods as well as develop new methods to study the quantum conductance problem.